Alex Russell

Developer-authored JavaScript is definitely a problem. Third party scripts and pixels (for analytics and tracking), and bad ads seem to be significant issues as well. Thinking browser makers can help here.

I'd stop harping on first-party code and focus on 3p code if I started seeing partners come to us with functional 1p experiences. Thus far, 1p is enough to sink you if you're doing "modern frontend development".

First-party seems like a much harder problem. You're fighting against the js ecosystem, how companies budget/ hire/ prioritize, how developers are trained, etc...

First-party *should* be easier: it's a straightforward agency issue. Promote developers on a site-performance-weighted basis. You launched a new feature? Great! You made the site faster in a growth market? TAKE MY MONEY.

You're making assumptions about incentive structures in development organizations and that growth markets are priorities for all companies.

My experience also tells me your arguments carryore weight in *engineering* orgs. But not *marketing* orgs. Web things built by agencies/ vendors are often built to please stakeholders, not users. The problem is (at least) as much cultural as technical.

Definitionally, growth is interesting to businesses. I'm not saying that your growth market has to be BRIC countries, just that the org needs to shift to pay for performance (literally) in whatever markets you *do* care about.

I mean software in general has kept up with consuming every win from moore's law for like 4 decades, so it seems like in general software developers are bad.

The web has gone super-linear, tho. Our data shows we're getting worse faster than HW is improving. Also, in many segments, HW is not improving. Double-whammy, and JS bloat is the root cause.

I'd be curious for some analysis on what bits are growing, if you understand what I'm asking. Like, as in, what is it that is trying to be done that we could avoid here with better features and still meet expectations

The community is largely rejecting non-user-space answers. Many causes, but all lead to bad client-side experiences.

Sorry, can you clarify? Really don't know what you mean there

Rephrased: what's an example of a thing lots of people are using that is a significant chunk and has a native answer that isn't getting used? Is that what you mean?

Really i would like to better understand what you're saying ^^^

Ah, sorry for the late reply. One small example: I can't tell you how many copies of Promise (and Promise-alike) polyfills I see in apps I dissect.

You've been beating this drum for years, things haven't changed. It's a tough spot to be in, trying to get software developers to write less software. Hackers gonna hack.

It's true that things don't have to continuously improve. Won't keep me from trying, tho.

What I mean is that I wish there was some way to use people's natural inclinations to work toward your goal, rather than working against them.

This is a pricing problem -- a market failure in which developers externalize their costs. A good way for this to happen would be for sites to run 5-10x slower when devtools is open, but I lost that argument with Pavel and @paul_irish years go.

Exactly. And in a market environment Paul's argument is what's going to happen. I think that your other (dogged, indefatigable) efforts - to get pwas first class on mobile - is the best way to fix this. Once there is a big enough benefit the market should discover it and target